package ‘shinyitemanalysis’ - r · a datamedical is a data.frame consisting of 2,392...

Package ‘ShinyItemAnalysis’January 28, 2020

Type Package

Title Test and Item Analysis via Shiny

Version 1.3.2

Date 2020-01-27

Author Patricia Martinkova [aut, cre],Adela Hladka [aut],Ondrej Leder [ctb],Jakub Houdek [ctb],Lubomir Stepanek [ctb],Tomas Jurica [ctb],Jana Vorlickova [ctb],Jan Netik [ctb]

Maintainer Patricia Martinkova <[email protected]>

Depends R (>= 3.5.0)

Imports corrplot, cowplot, CTT, data.table, deltaPlotR, DT, difNLR (>=1.2.2), difR (>= 5.0), ggdendro, ggplot2 (>= 2.2.1), gridExtra,knitr, latticeExtra, ltm, mirt (>= 1.24), moments, msm, nnet,plotly, psych, psychometric, reshape2, rmarkdown, shiny (>=1.0.3), shinyBS, shinydashboard, shinyjs (>= 0.9), stringr,VGAM, xtable

Description Interactive shiny application for analysis of educational tests andtheir items.

License GPL-3

LazyData TRUE

RoxygenNote 7.0.2

BugReports https://github.com/patriciamar/ShinyItemAnalysis/issues

Encoding UTF-8

NeedsCompilation no

Repository CRAN

Date/Publication 2020-01-28 20:50:02 UTC

1

https://github.com/patriciamar/ShinyItemAnalysis/issues

2 dataMedical

R topics documented:dataMedical . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2dataMedicalgraded . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3dataMedicalkey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4dataMedicaltest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5DDplot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6DistractorAnalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9gDiscrim . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10ggWrightMap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12HCI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14HCIkey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15HCItest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15ItemAnalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16LearningToLearn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19plotAdjacent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20plotCumulative . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22plotDIFirt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23plotDIFLogistic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24plotDistractorAnalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26plotMultinomial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28startShinyItemAnalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29theme_app . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

Index 31

dataMedical Dichotomous Dataset of Admission Test to Medical School

Description

The dataMedical dataset consists of the responses of 2,392 subjects (750 males, 1,633 females and9 subjects without gender specification) to admission test to a medical school. It contains 100 items.A correct answer is coded as "1" and incorrect answer as "0". Missing answers were evaluated asincorrect, i.e. "0".

Usage

data(dataMedical)

Format

A dataMedical is a data.frame consisting of 2,392 observations on the following 102 variables.

X the first 100 columns represent dichotomously scored items of the test.

gender variable describing gender; values "0" and "1" refer to males and females.

StudySuccess criterion variable; value "1" means that student studies standardly, "0" otherwise(e.g. leaving or interrupting studies).

dataMedicalgraded 3

Author(s)

Cestmir StukaFirst Faculty of Medicine, Charles University

Martin VejrazkaFirst Faculty of Medicine, Charles University

Patricia MartinkovaInstitute of Computer Science of the Czech Academy of Sciences<[email protected]>

References

Stuka, C. Vejrazka, M., Martinkova, P. Komenda, M. & Stepanek, L. (2016). The Use of Test andItem Analysis for Improvement of Tests. Workshop held at conference MEFANET, 2016, Brno,Czech Republic.

See Also

dataMedicaltest, dataMedicalkey, dataMedicalgraded

dataMedicalgraded Graded Dataset of Admission Test to Medical School

Description

The dataMedicalgraded dataset consists of the responses of 2,392 subjects (750 males, 1,633females and 9 subjects without gender specification) to multiple-choice admission test to a medicalschool. It contains 100 items. Each item is graded with 0 to 4 points. Maximum of 4 points wereset if all correct answers and none of incorrect answers were selected.

Usage

data(dataMedicalgraded)

Format

A dataMedicalgraded is a data.frame consisting of 2,392 observations on the following 102variables.

X the first 100 columns represent ordinal item scores of the test.

gender variable describing gender; values "0" and "1" refer to males and females.

StudySuccess criterion variable; value "1" means that student studies standardly, "0" otherwise(e.g. leaving or interrupting studies).

4 dataMedicalkey

Author(s)




References

Stuka, C. Vejrazka, M., Martinkova, P. Komenda, M. & Stepanek, L. (2016). The Use of Test andItem Analisis for Improvement of Tests. Workshop held at conference MEFANET, 2016, Brno,Czech Republic.

See Also

dataMedical, dataMedicaltest, dataMedicalkey

dataMedicalkey Key of Correct Answers for dataMedicaltest Dataset

Description

The dataMedicalkey is a vector of factors representing correct answers of dataMedicaltestdataset.

Usage

data(dataMedicalkey)

Format

A vector with 100 values representing correct answers to items of dataMedicaltest dataset. Formore details see dataMedicaltest.

Author(s)



dataMedicaltest 5


References


See Also

dataMedical, dataMedicaltest, dataMedicalgraded

dataMedicaltest Dataset of Admission Test to Medical School

Description

The dataMedicaltest dataset consists of the responses of 2,392 subjects (750 males, 1,633 femalesand 9 subjects without gender specification) to multiple-choice admission test to a medical school.It contains 100 items, possible answers were A, B, C, D, while any combination of these can becorrect.

Usage

data(dataMedicaltest)

Format

A ‘dataMedicaltest‘ is a ‘data.frame‘ consisting of 2,392 observations on the following 102 vari-ables.

X the first 100 columns represent items answers.

gender variable describing gender; values ‘"0"‘ and ‘"1"‘ refer to males and females.

StudySuccess criterion variable; value ‘"1"‘ means that student studies standardly, ‘"0"‘ otherwise(e.g. leaving or interrupting studies).

Author(s)



6 DDplot

Patricia MartinkovaInstitute of Computer Science of the Czech Academy of [email protected]

References


See Also

dataMedical, dataMedicalkey, dataMedicalgraded

DDplot Graphical representation of difficulty and (generalized) discriminationin item analysis

Description

Plots difficulty and (generalized) discrimination for items ordered by difficulty.

Usage

DDplot(data, item.names, k = 3, l = 1, u = 3,discrim = "ULI", maxscore, minscore, bin = FALSE, cutscore, average.score = FALSE,thr = 0.2)

Arguments

data numeric: binary or ordinal data matrix or data frame. See Details.

item.names character: the names of items.

k numeric: number of groups to which may be data.frame x divided by the totalscore. Default value is 3. See Details.

l numeric: lower group. Default value is 1. See Details.

u numeric: upper group. Default value is 3. See Details.

discrim character: type of discrimination index to be calculated. Deafult value is "ULI".See Details.

maxscore vector or numeric: maximal scores of items. If numeric, the same maximal scoreis used for all items. If missing, vector of achieved maximal scores is calculatedand used in calculations.

minscore vector or numeric: minimal scores of items. If numeric, the same minimal scoreis used for all items. If missing, vector of achieved minimal scores is calculatedand used in calculations.

DDplot 7

bin logical: should the ordinal data be binarized. Deafult value is FALSE. See De-tails.

cutscore vector or numeric: cutscore used to binarize the data.set. If numeric, the samecutscore is used for all items. If missing, vector of maximal scores is used incalculations.

average.score logical: should average score of the item disaplyed instead of difficulty. Defaultvalue is FALSE. See Details.

thr numeric: value of discrimination threshold. See Details.

Details

The data is a matrix or data frame whose rows represents examinee answers (1 correct, 0 incorrect,or ordinal item scores) and columns correspond to the items.

The item.names argument stands for names of items. If not specified, the names of dataset columnsare used. Difficulty and discrimination indices are plotted for each item, items are ordered by theirdifficulty.

Discrimination is calculated using method specified in discrim. Default option "ULI" calculatesdifference in ratio of correct answers in upper and lower third of students. "RIT" index caluclatescorrelation between item score and test total score. "RIR" index caclulates correlation between itemscore and total score for the rest of the items. With option "none", only difficulty is displayed.

"ULI" index can be generalized using arguments k, l and u. Generalized ULI discrimination isthen computed as follows: The function takes data on individuals, computes their total test scoreand then divides individuals into k groups. The lower and upper group are determined by l and uparameters, i.e. l-th and u-th group where the ordering is defined by increasing total score.

For ordinal data, difficulty is defined as relative score (achieved - minimal)/(maximal - minimal).Minimal score can be specified by minscore, maximal score can be specified by maxscore. Aver-age score of items can be displayed with argument average.score = T. Note that for binary datadifficulty estimate is the same as average score of the item.

Binarization of data is allowed in bin, for this purpose cutscore is used.

By rule of thums, discrimination of items should not be lower than 0.2. The value of threshold canbe specified via thr argument. In case that thr = NULL, no horizontal line is displayed in the plot.

Note

Generalized discrimination is calculated by gDiscrim function.

Author(s)

Adela HladkaInstitute of Computer Science of the Czech Academy of SciencesFaculty of Mathematics and Physics, Charles University<[email protected]>

Lubomir StepanekFirst Faculty of Medicine, Charles University

8 DDplot

Jana VorlickovaInstitute of Computer Science of the Czech Academy of Sciences


References

Martinkova, P., Stepanek, L., Drabinova, A., Houdek, J., Vejrazka, M., & Stuka, C. (2017). Semi-real-time analyses of item characteristics for medical school admission tests. In: Proceedings of the2017 Federated Conference on Computer Science and Information Systems.

See Also

gDiscrim, discrim

Examples

# loading 100-item medical admission test data setsdata(dataMedical, dataMedicalgraded)# binary data setdataBin <- dataMedical[, 1:100]# ordinal data setdataOrd <- dataMedicalgraded[, 1:100]

# DDplot of binary data setDDplot(dataBin)## Not run:#' # DDplot of binary data set without thresholdDDplot(dataBin, thr = NULL)# compared to DDplot using ordinal data set and 'bin = TRUE'DDplot(dataOrd, bin = TRUE)# compared to binarized data set using bin = TRUE and cutscore equal to 3DDplot(dataOrd, bin = TRUE, cutscore = 3)

# DDplot of binary data using generalized ULI# discrimination based on 5 groups, comparing 4th and 5th# threshold lowered to 0.1DDplot(dataBin, k = 5, l = 4, u = 5, thr = 0.1)

# DDplot of ordinal data set using ULIDDplot(dataOrd)# DDplot of ordinal data set using generalized ULI# discrimination based on 5 groups, comparing 4th and 5th# threshold lowered to 0.1DDplot(dataOrd, k = 5, l = 4, u = 5, thr = 0.1)# DDplot of ordinal data set using RITDDplot(dataOrd, discrim = "RIT")# DDplot of ordinal data set using RIR

DistractorAnalysis 9

DDplot(dataOrd, discrim = "RIR")# DDplot of ordinal data set disaplaying only difficultyDDplot(dataBin, discrim = "none")

# DDplot of ordinal data set disaplaying difficulty estimatesDDplot(dataOrd)# DDplot of ordinal data set disaplaying average item scoresDDplot(dataOrd, average.score = TRUE)

## End(Not run)

DistractorAnalysis Function for item distractor analysis

Description

Performs distractor analysis for each item and optional number of groups.

Usage

DistractorAnalysis(data, key, p.table = FALSE, num.groups = 3, matching = NULL,match.discrete = FALSE, cut.points)

Arguments

data character: data matrix or data frame. See Details.

key character: answer key for the items.

p.table logical: should the function return the proportions. If FALSE (default) the countsare returned.

num.groups numeric: number of groups to that should be respondents splitted.

matching numeric: numeric vector. If not provided, total score is calculated and distractoranalysis is performed based on it.

match.discrete logical: is matching discrete? Default value is FALSE. See details.

cut.points numeric: numeric vector specifying cut points of matching. See details.

Details

This function is adapted version of distractor.analysis function from CTT package.

The data is a matrix or data frame whose rows represents unscored item response from a multiple-choice test and columns correspond to the items.

The key must be a vector of the same length as ncol(data).

In case, no matching is provided, the scores are calculated using the item data and key. Therespondents are by default splitted into the num.groups-quantiles and the number (or proportion)of respondents in each quantile is reported with respect to their answers. In case that matchingis discrete (match.discrete = TRUE), matching is splitted based on its unique levels. Other cutpoints can be specified via cut.points argument.

10 gDiscrim

Author(s)



See Also

distractor.analysis

Examples

# loading 100-item medical admission test datadata(dataMedicaltest, dataMedicalkey)data <- dataMedicaltest[, 1:100]dataBin <- dataMedical[, 1:100]key <- unlist(dataMedicalkey)

# distractor analysis for dataMedicaltest data setDistractorAnalysis(data, key)## Not run:# distractor analysis for dataMedicaltest data set with proportionsDistractorAnalysis(data, key, p.table = T)

# distractor analysis for dataMedicaltest data set for 6 groupsDistractorAnalysis(data, key, num.group = 6)

# distractor analysis for dataMedicaltest using specified matchingmatching <- round(rowSums(databin), -1)DistractorAnalysis(data, key, matching = matching)

# distractor analysis for dataMedicaltest using discrete matchingDistractorAnalysis(data, key, matching = matching, match.discrete = T)

# distractor analysis for dataMedicaltest using groups specified by cut.pointsDistractorAnalysis(data, key, cut.points = seq(10, 100, 10))

## End(Not run)

gDiscrim Generalized Item Discrimination

gDiscrim 11

Description

gDiscrim function computes various generalizations of discrimination index ULI. It enumerates theablitity of item to distinguish between individuals from upper (U) vs. lower (L) ability groups, i.e.between respondents with high vs. low overall score on the test. Number of groups, as well as upperand lower groups can be specified by user. Maximal and minimal score in ordinal data sets can bespecified by user.

Usage

gDiscrim(x, k = 3, l = 1, u = 3, maxscore, minscore)

Arguments

x matrix or data.frame of items to be examined. Rows represent persons, columnsreperesent items.




maxscore numeric: maximal score in ordinal items. If missing, vector of obtained maximalscores is imputed. See Details.

minscore numeric: minimal score in ordinal items. If missing, vector of obtained minimalscores is imputed. See Details.

Details

The function computes total test scores for all respondents and then divides the respondents into kgroups. The lower and upper groups are determined by l and u parameters, i.e. l-th and u-th groupwhere the ordering is defined by increasing total score.

In ordinal items, difficulty is calculated as difference of average score divided by range (maximalpossible score maxscore minus minimal possible score minscore for given item).

Discrimination is calculated as difference in difficulty between upper and lower group.

Note

gDiscrim is used by DDplot function.

Author(s)


Lubomir StepanekFirst Faculty of Medicine, Charles University

12 ggWrightMap



References

Martinkova, P., Stepanek, L., Drabinova, A., Houdek, J., Vejrazka, M., & Stuka, C. (2017). Semi-real-time analyses of item characteristics for medical school admission tests. In: Proceedings of the2017 Federated Conference on Computer Science and Information Systems. https://doi.org/10.15439/2017F380

See Also

DDplot

Examples

# loading 100-item medical admission test data setsdata(dataMedical, dataMedicalgraded)# binary data setdataBin <- dataMedical[, 1:100]# ordinal data setdataOrd <- dataMedicalgraded[, 1:100]

# ULI for first 5 items for binary data set# compare to psychometric::discrim(x)gDiscrim(dataBin)[1:5]# generalized ULI using 5 groups, compare 4th and 5th for binary data setgDiscrim(dataBin, k = 5, l = 4, u = 5)[1:5]

# ULI for first 5 items for ordinal data setgDiscrim(dataOrd)[1:5]# generalized ULI using 5 groups, compare 4th and 5th for binary data setgDiscrim(dataOrd, k = 5, l = 4, u = 5)[1:5]# maximum (4) and minimum (0) score are same for all itemsgDiscrim(dataOrd, k = 5, l = 4, u = 5, maxscore = 4, minscore = 0)[1:5]

ggWrightMap Wright Map using ggplot

Description

This function allows to generate Wright Map (also called item-person map) using ggplot functionfrom package ggplot2 and plot_grid function from cowplot. Wright Map is used to displayhistogram of factor scores and the item difficulty parameters estimated by the Rasch IRT model.

ggWrightMap 13

Usage

ggWrightMap(theta, b, binwidth = 0.5, color = "blue", size = 15, item.names)

Arguments

theta numeric: vector of ability estimates.

b numeric: vector of difficulty estimates.

binwidth numeric: the width of the bins of histogram.

color character: color of histogram.

size text size in pts.

item.names names of items to be displayed.

Author(s)



References

Wright, B. D., & Stone, M. H. (1979). Best test design.

See Also

wrightMap

Examples

library(mirt)

# loading 100-item medical admission test data setsdata(dataMedical)# binary data setdataBin <- dataMedical[, 1:100]

# fit Rasch model with mirt packagefit <- mirt(dataBin, model = 1, itemtype = "Rasch")# factor scorestheta <- as.vector(fscores(fit))# difficulty estimatesb <- coef(fit, simplify = TRUE)$items[, "d"]

14 HCI

ggWrightMap(theta, b)

item.names <- paste("Item", 1:20)ggWrightMap(theta, b, item.names = item.names)

HCI Homeostasis Concept Inventory Dichotomous Dataset

Description

(HCI) dataset consists of the dichotomously scored responses of 651 students (405 males, 246 fe-males) to Homeostasis Concept Inventory (HCI) multiple-choice test. It containts 20 items, vectorof gender membership and identificator whether students plan to major.

Usage

data(HCI)

Format

HCI is a data.frame consisting of 651 observations on the 22 variables.

Item1-Item20 dichotomously scored items of the HCI test.

gender gender membership vector, "0" males, "1" females.

major identificator whether students planning to major in the life sciences.

Author(s)

Jenny L. McFarlandBiology Department, Edmonds Community College

References

McFarland, J. L., Price, R. M., Wenderoth, M. P., Martinkova, P., Cliff, W., Michael, J., ... &Wright, A. (2017). Development and validation of the homeostasis concept inventory. CBE-LifeSciences Education, 16(2), ar35.

See Also

HCItest, HCIkey

HCIkey 15

HCIkey Key of Correct Answers for Homeostasis Concept Inventory Dataset

Description

The HCIkey is a vector of factors representing correct answers of HCItest dataset.

Usage

data(HCIkey)

Format

A nominal vector with 20 values representing correct answers to items of HCItest dataset. Formore details see HCItest.

Author(s)


References


See Also

HCI, HCItest

HCItest Homeostasis Concept Inventory Dataset

Description

(HCItest) dataset consists of the responses of 651 students (405 males, 246 females) to Homeosta-sis Concept Inventory (HCI) multiple-choice test. It containts 20 items, vector of gender member-ship and identificator whether students plan to major.

Usage

data(HCItest)

16 ItemAnalysis

Format

HCItest is a data.frame consisting of 651 observations on the 22 variables.

Item1-Item20 multiple-choice items of the HCI test.

gender gender membership vector, "0" males, "1" females.

major identificator whether students planning to major in the life sciences.

Author(s)


References


See Also

HCI, HCIkey

ItemAnalysis Item Analysis

Description

ItemAnalysis function computes various traditional item analysis indices including difficulty, dis-crimination and item validity. For ordinal items the difficulty and discrimination indices take intoaccount minimal item score as well as range.

Usage

ItemAnalysis(data, y = NULL, k = 3, l = 1, u = 3,maxscore, minscore, cutscore, add.bin = FALSE)

Arguments

data matrix or data.frame of items to be examined. Rows represent respondents,columns reperesent items.

y vector of criterion values.




ItemAnalysis 17

maxscore numeric or vector: maximal score in ordinal items. If missing, vector of obtainedmaximal scores is imputed. See Details.

minscore numeric or vector: minimal score in ordinal items. If missing, vector of obtainedminimal scores is imputed. See Details.

cutscore numeric or vector: cut score used for binarization of ordinal data. If missing,vector of maximal scores is imputed. See Details.

add.bin logical: If TRUE, indices are printed also for binarized data. See Details.

Details

For ordinal items the difficulty and discrimination indices take into account minimal item score aswell as range.

For calculation of discimination ULI index, it is possible to specify the number of groups k, andwhich two groups l and u are to be compared.

In ordinal items, difficulty is calculated as difference of average score divided by range (maximalpossible score maxscore minus minimal possible score minscore).

If add.bin is set to TRUE, item analysis of binarized data is included in the output table. In sucha case, cutscore is used for binarization. When binarizing the data, values greater or equal tocut-score are set to 1, other values are set to 0.

Value

ItemAnalysis function computes various traditional item analysis indices. Output is a data.framewith following columns:

Difficulty average score of the item divided by its range

Average score average score of the item

SD standard deviation of the item score

SD bin standard deviation of the item score for binarized dataCorrect answers

proportion of correct answers

Min score minimal score specified in minscore; if not provided, observed minimal score

Max score maximal score specified in maxscore; if not provided, observed maximal score

Obtained min observed minimal score

Obtained max observed maximal score

Cut score cut-score specified in cutscore

ULI generalized ULI

ULI default dscrimination with ULI

RIT correlation between item score and overall test score

RIR correlation between item score and overall test score

Item criterion correlation of item score with criterionItem reliability

item reliability index calculated as cor(item,test)*sqrt(((N-1)/N)*var(item)),see Allen & Yen (1979), Ch.6.4

18 ItemAnalysis

Item reliability woi

item reliability index (scored without item)

Item validity item validity index calculated as cor(item,y)*sqrt(((N-1)/N)*var(item)),see Allen & Yen (1979), Ch.6.4

Item criterion correlation between item and criterion y

Alpha drop Cronbach’s alpha without given item

With add.bin == TRUE, indices based on binarized data set are also provided and marked with bin.

Author(s)




References

Martinkova, P., Stepanek, L., Drabinova, A., Houdek, J., Vejrazka, M., & Stuka, C. (2017). Semi-real-time analyses of item characteristics for medical school admission tests. In: Proceedings of the2017 Federated Conference on Computer Science and Information Systems. https://doi.org/10.15439/2017F380

Allen, M. J. & Yen, W. M. (1979). Introduction to measurement theory. Monterey, CA: Brooks/Cole.

See Also

DDplot, gDiscrim

Examples

## Not run:# loading 100-item medical admission test data setsdata(dataMedical, dataMedicalgraded)# binary data setdataBin <- dataMedical[, 1:100]# ordinal data setdataOrd <- dataMedicalgraded[, 1:100]# study success is the same for both data setsStudySuccess <- dataMedical[, 102]

# item analysis for binary datahead(ItemAnalysis(dataBin))

LearningToLearn 19

# item analysis for binary data using also study successhead(ItemAnalysis(dataBin, y = StudySuccess))

# item analysis for binary datahead(ItemAnalysis(dataOrd))# item analysis for binary data using also study successhead(ItemAnalysis(dataOrd, y = StudySuccess))# including also item analysis for binarized datahead(ItemAnalysis(dataOrd,

y = StudySuccess, k = 5, l = 4, u = 5,maxscore = 4, minscore = 0, cutscore = 4, add.bin = TRUE

))

## End(Not run)

LearningToLearn Dichotomous Data Set of Learning to Learn Test

Description

LearningToLearn is a real longitudinal dataset used in Martinkova et al (2020) study, demonstrat-ing differential item functioning in change (DIF-C) on Learning to Learn (LtL) test. Among othervariables, it primarily contains binary-coded responses of 782 subjects to (mostly) multiple-choicetest consisting of 41 items within 7 subscales (see Format for details). Each respondent was testedtwice in total – the first time in Grade 6 and the second time in Grade 9. Most importantly, schooltrack (variable track_01 or track) is available, with 391 students attending basic school (BS)and 391 pursuing selective academic school (AS). This dataset was created using propensity scorematching algorithm to achieve similar characteristics in both tracks (see References for details). Tofurther simplify the work with LtL dataset, we provide computed total scores as well as 7 subscores,both for Grade 6 and Grade 9. The dataset also includes change variables for each item (see Formatfor details) for more detailed DIF-C analysis using multinomial regression model.

Usage

data(LearningToLearn)

Format

A LearningToLearn data frame consists of 782 observations on the following 141 variables:

track_01 dichotomously scored school track, where "1" denotes the selective academic schoolone.

track school track, where "AS" represents the selective academic school track, and "BS" stands forbasic school track.

score_6 & score_9 total test score value obtained by summing all 41 items of LtL, the numberdenotes the Grade which the respondent was taking at the time of testing.

score_6_subtest1–score_6_subtest7 scores of respective cognitive subtest (1–7) of LtL in Grade6.

20 plotAdjacent

score_9_subtest1–score_9_subtest7 scores of respective cognitive subtest (1–7) of LtL in Grade9.

Item1A_6–Item7F_6 dichotomously coded 41 individual items obtained at Grade 6, "1" repre-sents the correct answer to the particular item.

Item1A_9–Item7F_9 dichotomously coded 41 individual items obtained at Grade 9, "1" repre-sents the correct answer to the particular item.

Item1A_changes–Item7F_changes change patterns with those possible values:

• a student responded correctly in neither Grade 6 nor in Grade 9 (did not improve, "00")• a student responded correctly in Grade 6 but not in Grade 9 (deteriorated, "10")• a student did not respond correctly in Grade 6 but responded correctly in Grade 9 (im-

proved, "01"), and• a student responded correctly in both grades (did not deteriorate, "11")

Author(s)

Patricia MartinkovaFaculty of Education, Charles UniversityInstitute of Computer Science of the Czech Academy of Sciences<[email protected]>


Eva PotuznikovaFaculty of Education, Charles University

References

Martinkova, P., Hladka, A., & Potuznikova, E. (2020). Is academic tracking related to gains inlearning competence? Using propensity score matching and differential item change functioninganalysis for better understanding of tracking implications. Learning and Instruction, 66, 101286.https://doi.org/10.1016/j.learninstruc.2019.101286

plotAdjacent Function for plotting category probabilities of adjacent logistic regres-sion model

Description

Function for plotting category probabilities function estimated by vglm() from VGAM package

https://doi.org/10.1016/j.learninstruc.2019.101286

plotAdjacent 21

Usage

plotAdjacent(x, matching.name = "matching")

Arguments

x object of class vglm

matching.name character: name of matching criterion used for estimation in x.

Author(s)

Tomas Jurica



See Also

vglm

Examples

# loading packageslibrary(VGAM)

# loading datadata <- dataMedicalgraded[, 1:100]

# total score calculationscore <- apply(data, 1, sum)data[, 1] <- ordered(factor(data[, 1], levels = 0:max(data[, 1])))

# cummulative logistic model for item 1fit <- vglm(data[, 1] ~ score, family = acat(reverse = FALSE, parallel = TRUE))# coefficients for item 1coefs <- coef(fit)

plotAdjacent(fit, matching.name = "Total score")

22 plotCumulative

plotCumulative Function for plotting cumulative and category probabilities of cumu-lative logistic regression model

Description

Function for plotting cumulative and category probabilities function estimated by vglm() fromVGAM package

Usage

plotCumulative(x, type = "cumulative", matching.name = "matching")

Arguments

x object of class vglm

type character: type of plot to be displayed. Options are "cumulative" (default) forcumulative probabilities and "category" for category probabilities.


Author(s)

Tomas Jurica



See Also

vglm

Examples

# loading packageslibrary(VGAM)

# loading datadata <- dataMedicalgraded[, 1:100]

# total score calculation

plotDIFirt 23

score <- apply(data, 1, sum)data[, 1] <- ordered(factor(data[, 1], levels = 0:max(data[, 1])))

# cummulative logistic model for item 1fit <- vglm(data[, 1] ~ score, family = cumulative(reverse = TRUE, parallel = TRUE))# coefficients for item 1coefs <- coef(fit)

plotCumulative(fit, type = "cumulative", matching.name = "Total score")plotCumulative(fit, type = "category", matching.name = "Total score")

plotDIFirt Function for characteristic curve of DIF IRT model

Description

Plots characteristic curve of IRT model.

Usage

plotDIFirt(parameters, test = "Lord", item = "all", item.name, same.scale = F)

Arguments

parameters numeric: data matrix or data frame. See Details.

test character: type of statistic to be shown. See Details.

item either character ("all"), or numeric vector, or single number corresponding tocolumn indicators. See Details.

item.name character: the name of item.

same.scale logical: are the item parameters on the same scale? (default is "FALSE"). SeeDetails.

Details

This function plots characteristic curve of DIF IRT model.

The parameters matrix has a number of rows equal to twice the number of items in the data set.The first J rows refer to the item parameter estimates in the reference group, while the last J onescorrespond to the same items in the focal group. The number of columns depends on the selectedIRT model: 2 for the 1PL model, 5 for the 2PL model, 6 for the constrained 3PL model and 9 forthe unconstrained 3PL model. The columns of irtParam have to follow the same structure as theoutput of itemParEst, difLord or difRaju command from difR package.

Two possible type of test statistics can be visualized - "Lord" gives only characteristic curves,"Raju" also highlights area between these curves.

For default option "all", all characteristic curves are plotted.

24 plotDIFLogistic

Author(s)



See Also

itemParEst, difLord, difRaju

Examples

# loading librarieslibrary(difR)library(ltm)

# loading data based on GMAT2data(GMAT2, package = "difNLR")

# Estimation of 2PL IRT model and Lord's statistic# by difR packagefitLord <- difLord(GMAT2, group = 21, focal.name = 1, model = "2PL")# plot of item 1 and Lord's statisticplotDIFirt(fitLord$itemParInit, item = 1)

# Estimation of 2PL IRT model and Raju's statistic# by difR packagefitRaju <- difRaju(GMAT2, group = 21, focal.name = 1, model = "2PL")# plot of item 1 and Lord's statisticplotDIFirt(fitRaju$itemParInit, test = "Raju", item = 1)

plotDIFLogistic Function for characteristic curve of 2PL logistic DIF model

Description

Plots characteristic curve of 2PL logistic DIF model

Usage

plotDIFLogistic(x, item = 1, item.name, group.names = c("Reference", "Focal"),Data, group, match, draw.empirical = TRUE)

plotDIFLogistic 25

Arguments

x an object of "Logistic" class. See Details.

item numeric: number of item to be plotted

item.name character: the name of item to be used as title of plot.

group.names character: names of reference and focal group.

Data numeric: the data matrix. See Details.

group numeric: the vector of group membership. See Details.

match character or numeric: specifies matching criterion. Can be either "score", ornumeric vector of the same length as number of observations in Data. See De-tails.

draw.empirical logical: whether empirical probabilities should be calculated and plotted. De-fault value is TRUE.

Details

This function plots characteristic curves of 2PL logistic DIF model fitted by difLogistic() func-tion from difR package using ggplot2.

Data and group are used to calculate empirical probabilities for reference and focal group. matchshould be the same as in x$match. In case that matching variable is used instead of total score orstandardized score, match needs to be a numeric vector of the same the same length as number ofobservations in Data.

Author(s)



See Also

difLogistic, ggplot

Examples

# loading librarieslibrary(difR)

# loading data based on GMATdata(GMAT, package = "difNLR")Data <- GMAT[, 1:20]group <- GMAT[, 21]

26 plotDistractorAnalysis

# DIF detection using difLogistic() functionx <- difLogistic(Data, group, focal.name = 1)# Characteristic curve by logistic regression modelplotDIFLogistic(x, item = 1, Data = Data, group = group)

# Using name of column as item identifierplotDIFLogistic(x, item = "Item1", Data = Data, group = group)

# Renaming reference and focal groupplotDIFLogistic(x, item = 1, group.names = c("Group 1", "Group 2"), Data = Data, group = group)

# Not plotting empirical probabilitiesplotDIFLogistic(x, item = 1, draw.empirical = FALSE)

plotDistractorAnalysis

Function for graphical representation of item distractor analysis

Description

Plots graphical representation of item distractor analysis with proportions and optional number ofgroups.

Usage

plotDistractorAnalysis(data, key, num.groups = 3, item = 1, item.name,multiple.answers = TRUE, matching = NULL, match.discrete = FALSE, cut.points)

Arguments

data character: data matrix or data frame. See Details.

key character: answer key for the items.

num.groups numeric: number of groups to that should be respondents splitted.

item numeric: the number of item to be plotted.

item.name character: the name of item.multiple.answers

logical: should be all combinations plotted (default) or should be answers split-ted into distractors. See Details.

matching numeric: numeric vector. If not provided, total score is calculated and distractoranalysis is performed based on it.

match.discrete logical: is matching discrete? Default value is FALSE. See details.

cut.points numeric: numeric vector specifying cut points of matching. See details.

plotDistractorAnalysis 27

Details

This function is graphical representation of DistractorAnalysis function. In case, no matchingis provided, the scores are calculated using the item data and key. The respondents are by defaultsplitted into the num.groups-quantiles and the proportions of respondents in each quantile are dis-played with respect to their answers. In case that matching is discrete (match.discrete = TRUE),matching is splitted based on its unique levels. Other cut points can be specified via cut.pointsargument.

The data is a matrix or data frame whose rows represents unscored item response from a multiple-choice test and columns correspond to the items.

The key must be a vector of the same length as ncol(data). In case it is not provided, matchingneed to be specified.

If multiple.answers = TRUE (default) all reported combinations of answers are plotted. If multiple.answers= FALSE all combinations are splitted into distractors and only these are then plotted with correctcombination.

Author(s)



See Also

DistractorAnalysis, distractor.analysis

Examples

# loading 100-item medical admission test datadata(dataMedical, dataMedicaltest, dataMedicalkey)data <- dataMedicaltest[, 1:100]dataBin <- dataMedical[, 1:100]key <- unlist(dataMedicalkey)

# distractor plot for items 48, 57 and 32 displaying distractors onlyplotDistractorAnalysis(data, key, item = 48, multiple.answers = FALSE)# correct answer B does not function wellplotDistractorAnalysis(data, key, item = 57, multiple.answers = FALSE)# all options function well, thus the whole item discriminates wellplotDistractorAnalysis(data, key, item = 32, multiple.answers = FALSE)# functions well, thus the whole item discriminates well## Not run:# distractor plot for items 48, 57 and 32 displaying all combinations

28 plotMultinomial

plotDistractorAnalysis(data, key, item = 48)plotDistractorAnalysis(data, key, item = 57)plotDistractorAnalysis(data, key, item = 32)

# distractor plot for item 57 with all combinations and 6 groupsplotDistractorAnalysis(data, key, item = 57, num.group = 6)

# distractor plot for item 57 using specified matching and key optionmatching <- round(rowSums(dataBin), -1)plotDistractorAnalysis(data, key, item = 57, matching = matching)# distractor plot for item 57 using specified matching without key optionplotDistractorAnalysis(data, item = 57, matching = matching)

# distractor plot for item 57 using discrete matchingplotDistractorAnalysis(data, key, item = 57, matching = matching, match.discrete = T)

# distractor plot for item 57 using groups specified by cut.pointsplotDistractorAnalysis(data, key, item = 57, cut.points = seq(10, 100, 10))

## End(Not run)

plotMultinomial Function for plotting category probabilities of multinomial log-linearregression model

Description

Plots category probabilities functions estimated by multinom() from nnet package.

Usage

plotMultinomial(x, matching, matching.name = "matching")

Arguments

x object of class multinom

matching numeric: vector of matching criterion used for estimation in x.


Author(s)


Tomas Jurica

startShinyItemAnalysis 29


See Also

multinom

Examples

# loading datadata(GMAT, GMATtest, GMATkey, package = "difNLR")

matching <- scale(apply(GMAT[, 1:20], 1, sum)) # Z-scoredata <- GMATtest[, 1:20]key <- GMATkey

# multinomial model for item 1fit <- nnet::multinom(relevel(data[, 1], ref = paste(key[1])) ~ matching)

# plotting category probabilitiesplotMultinomial(fit, matching, matching.name = "Z-score")

startShinyItemAnalysis

This function will start ShinyItemAnalysis application.

Description

An interactive shiny application for running test and item analysis.

Usage

startShinyItemAnalysis()

Author(s)



30 theme_app

Examples

## Not run:rm(list = ls())startShinyItemAnalysis()

## End(Not run)

theme_app Complete theme for ShinyItemAnalysis graphics

Description

This complete theme is based on theme_bw and it was modified for purposes of ShinyItemAnalysis.

Usage

theme_app(base_size = 15, base_family = "")

Arguments

base_size base font size

base_family base font family

See Also

ggtheme

Examples

library(ggplot2)data(GMAT, package = "difNLR")data <- GMAT[, 1:20]# total score calculationdf <- data.frame(score = apply(data, 1, sum))# histogramg <- ggplot(df, aes(score)) +

geom_histogram(binwidth = 1) +xlab("Total score") +ylab("Number of respondents")

gg + theme_app()

Index

∗Topic datasetsdataMedical, 2dataMedicalgraded, 3dataMedicalkey, 4dataMedicaltest, 5HCI, 14HCIkey, 15HCItest, 15LearningToLearn, 19

dataMedical, 2, 4–6dataMedicalgraded, 3, 3, 5, 6dataMedicalkey, 3, 4, 4, 6dataMedicaltest, 3–5, 5DDplot, 6, 11, 12, 18difLogistic, 25difLord, 24difRaju, 24discrim, 8distractor.analysis, 9, 10, 27DistractorAnalysis, 9, 27

gDiscrim, 7, 8, 10, 18ggplot, 25ggtheme, 30ggWrightMap, 12

HCI, 14, 15, 16HCIkey, 14, 15, 16HCItest, 14, 15, 15

ItemAnalysis, 16itemParEst, 24

LearningToLearn, 19

multinom, 29

plotAdjacent, 20plotCumulative, 22plotDIFirt, 23

plotDIFLogistic, 24plotDistractorAnalysis, 26plotMultinomial, 28

startShinyItemAnalysis, 29

theme_app, 30

vglm, 21, 22

wrightMap, 13

31

package ‘shinyitemanalysis’ - r · a datamedical is a data.frame consisting of 2,392...

Documents